Prediction-Oriented Segmentation Partial Least Square (POS-PLS) in The Case of Child Underutrition in East Java Fatkhi Rizqiyah Agustina, Bambang Widjanarko Otok, Santi Wulan Purnami
Department of Statistics, Faculty of Science and Data Analytics, Institut Teknologi Sepuluh Nopember, Kampus ITS -Sukolilo, Surabaya 60111, Indonesia
Abstract
Structural equation modeling (SEM) is a statistical technique that can explain relatively complex relationship structures involving many variables. This is a structural equation model. Fulfillment of parametric distribution assumptions is often difficult to fulfill, so Partial Least Square (SEM-PLS) is a good alternative to overcome these limitations. SEM-PLS assumes that the sample taken comes from a homogeneous population. Whereas in research, many variables and indicators were involved which were collected from populations with different characteristics resulting in heterogeneity in the data. There are two kinds of data heterogeneity, namely observed heterogeneity and unobserved heterogeneity. Ignoring unobserved heterogeneity will lead to some research errors. One method to overcome this is prediction-oriented segmentation PLS (POS-PLS). One aspect with unobserved heterogeneity is the health aspect, which is related to malnutrition. Inadequate nutrition or often known as malnutrition has a very broad impact, not only having a big role in increasing morbidity and mortality but also having a role in disrupting psychosocial aspects and intellectual development. Malnutrition can affect anyone, but infants and toddlers are the most vulnerable group because they require high levels of nutrients for growth and development. Stunting, wasting, and being underweight are expressions of a lack of energy and protein intake, infectious diseases, and also the result of malnutrition during pregnancy. Nutritional status is influenced by 3 factors, namely direct, basic, and enabling factors that formed into latent variables. The endogenous latent variables used were food, practice, and service. The exogenous latent variable is social economy.
